Research Square

Frequency-Specific Age Associations in EEG Microstate Dynamics Across the Adult Lifespan

Preprint: cross-sectional study of 60 peopleMechanisms
Follow

Abstract

EEG microstates are conventionally derived from broadband signals, although their temporal dynamics may differ across frequency ranges. Whether age associations in conventional voltage microstates are similarly frequency specific remains unclear. We examined resting-state EEG from 60 healthy adults aged 20–78 years and derived microstates separately from broadband (1–30 Hz), delta (1–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), and beta (15–30 Hz) activity. The number of microstates was selected independently for each band using the Krzanowski–Lai criterion, yielding four classes for broadband, delta, alpha, and beta, and five for theta. Age associations were tested for duration, occurrence, coverage, global explained variance, observed transition probabilities, and transitions adjusted for those expected from class prevalence, with false-discovery-rate correction. Age associations showed a marked spectral dependence. Of 84 state-level tests, 13 survived correction: one in broadband, six in delta, one in theta, none in alpha, and five in beta. In delta, increasing age was associated with greater expression of MS B and reduced expression of MS C across duration, occurrence, and coverage. In beta, age was associated with greater MS A expression and reduced MS C expression. All significant state-level associations remained after adjustment for sex, education, and band-specific spectral power. Eight observed transition associations survived correction, all in broadband and delta, but only three remained after accounting for transitions expected from state prevalence. These findings indicate that age-related variation in EEG microstate dynamics is strongly frequency-specific and is expressed primarily through redistribution of state expression rather than through widespread alteration of transition preference. Frequency-resolved microstate analysis, hence, reveals age associations that are not evident in conventional broadband analysis.